Taxable Investing > Complexities > Stabilizing Correlations

Post Modern Optimization refers to efforts to improve Mean Variance Optimization.  Empirical studies have indicated that MV optimization suffers from estimation error in the variance-covariance matrix or similarly in the correlation matrix due to large standard errors.  This estimation error leads to inaccurate asset allocation and correspondingly poor portfolio forecasting.

There have been a number of different methods developed to address this issue including Shrinkage Methods, Resampling, Black-Litterman and Random Matrix Theory (RMT).  PORTAX uses RMT for a number of different reasons that are discussed in the Help Documentation.  One key feature of RMT, is that it allows PORTAX users to either use historical correlations based on time series or use forecasted correlations without the need for time series.


PORTAX Correlation Filtering

Click on image to enlarge ...